Electronic Theses and Dissertation

Penelitian ini menganalisis pengaruh penerapan Intelligent Reflecting Surface (IRS) terhadap efisiensi energi sistem komunikasi 6G berbasis frekuensi terahertz dengan dukungan Machine Learning. IRS berfungsi sebagai permukaan reflektif pasif yang mampu mengarahkan gelombang elektromagnetik secara adaptif untuk memperkuat sinyal dan mengoptimalkan transmisi data. Tujuan penelitian ini adalah mengevaluasi pengaruh jumlah elemen IRS, bandwidth, jarak, dan daya Electromagnetic Interference (EMI) terhadap Signal-to-Noise Ratio (SNR), Achievable Rate (AR), dan Energy Efficiency (EE), serta menilai kemampuan model Random Forest (RF) dan Extreme Gradient Boosting (XGBoost) dalam memprediksi kinerja sistem. Metode penelitian dilakukan melalui simulasi numerik pada frekuensi 1 THz dengan variasi jumlah elemen IRS (25-1600), bandwidth (0,3-0,8 GHz), serta kondisi dengan EMI. Hasil penelitian menunjukkan bahwa peningkatan jumlah elemen secara signifikan meningkatkan nilai SNR dan EE, serta menjaga kestabilan kinerja sistem pada jarak transmisi yang lebih jauh. Selain itu, peningkatan bandwidth dari 0,3 GHz menjadi 0,8 GHz menaikkan AR dari 4 Gbps menjadi lebih dari 11 Gbps. Model XGBoost memberikan akurasi prediksi terbaik dengan nilai R² = 0,994 dan RMSE = 0.017. Secara keseluruhan, penerapan IRS terbukti mampu meningkatkan kualitas sinyal, kapasitas transmisi, dan efisiensi energi, sedangkan integrasi dengan Machine Learning (ML) memberikan kemampuan prediksi yang cepat dan akurat. Hasil ini menunjukkan bahwa kombinasi IRS dan Machine Learning berpotensi menjadi solusi kunci dalam pengembangan sistem komunikasi 6G yang efisien dan adaptif di masa mendatang.
Kata kunci: Intelligent Reflecting Surface, 6G, Signal-to-Noise Ratio, Achievable Rate, Energy Efficiency, Random Forest , XGBoost, Machine Learning.

This study analyzes the impact of implementing an Intelligent Reflecting Surface (IRS) on the energy efficiency of terahertz-frequency-based 6G communication systems supported by Machine Learning. IRS functions as a passive reflective surface capable of adaptively steering electromagnetic waves to enhance signal strength and optimize data transmission. The objective of this research is to evaluate the effects of the number of IRS elements, bandwidth, transmission distance, and Electromagnetic Interference (EMI) power on the Signal-to-Noise Ratio (SNR), Achievable Rate (AR), and Energy Efficiency (EE), as well as to assess the capability of Random Forest (RF) and Extreme Gradient Boosting (XGBoost) models in predicting system performance. The research methodology is conducted through numerical simulations at a frequency of 1 THz, with variations in the number of IRS elements (25–1600), bandwidth (0.3–0.8 GHz), and operating conditions in the presence of EMI. The results indicate that increasing the number of IRS elements significantly enhances SNR and EE, while maintaining stable system performance over longer transmission distances. In addition, increasing the bandwidth from 0.3 GHz to 0.8 GHz raises the AR from 4 Gbps to more than 11 Gbps. The XGBoost model achieves the best prediction accuracy, with an R² value of 0.994 and an RMSE of 0.017. Overall, the application of IRS is proven to improve signal quality, transmission capacity, and energy efficiency, while integration with Machine Learning (ML) enables fast and accurate performance prediction. These findings demonstrate that the combination of IRS and Machine Learning has strong potential as a key solution for the development of efficient and adaptive 6G communication systems in the future. Kata kunci: Intelligent Reflecting Surface, 6G, Signal-to-Noise Ratio, Achievable Rate, Energy Efficiency, Random Forest , XGBoost, Machine Learning